Jiajun Yang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Jiajun Yang has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
- Keyan Chen
Department of Aerospace Intelligent Science and Technology, School of Astronautics, Beihang University, Beijing, China
3 shared publications - Zhengxia Zou
Department of Aerospace Intelligent Science and Technology, School of Astronautics, Beihang University, Beijing, China
3 shared publications - Zhenwei Shi
Beihang University
3 shared publications - Liqin Liu
Beihang University
1 shared publication - Wenjing Wang
Image Processing Center, School of Astronautics, State Key Laboratory of Virtual Reality Technology and Systems, Beihang University, Beijing, China
1 shared publication Wenyuan Li
Department of Geography, University of Hong Kong, Hong Kong, China
1 shared publication
Latest Publications
- 12 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 4704316
CloudMamba: An Uncertainty-Guided Dual-Scale Mamba Network for Cloud Detection in Remote Sensing Imagery
Cloud detection in remote sensing imagery is a fundamental, critical, and highly challenging problem. Existing deep learning-based cloud detection methods generally formulate it as a single-stage pixel-wise binary segmentation task with one forward pass. However, such single-stage approaches exhibit ambiguity and uncertainty in thin-cloud regions and struggle to accurately handle fragmented clouds and boundary details. In this paper, we propose a novel deep learning framework termed CloudMamba. To address the ambiguity in thin-cloud regions, we introduce an uncertainty-guided two-stage cloud detection strategy. An embedded uncertainty estimation module is proposed to automatically quantify the confidence of thin-cloud segmentation, and a second-stage refinement segmentation is introduced to improve the accuracy in low-confidence hard regions. To better handle fragmented clouds and fine-grained boundary details, we design a dual-scale Mamba network based on a CNN-Mamba hybrid architecture. Compared with Transformer-based models with quadratic computational complexity, the proposed method maintains linear computational complexity while effectively capturing both large-scale structural characteristics and small-scale boundary details of clouds, enabling accurate delineation of overall cloud morphology and precise boundary segmentation. Extensive experiments conducted on the GF1_WHU and Levir_CS public datasets demonstrate that the proposed method outperforms existing approaches across multiple segmentation accuracy metrics, while offering high efficiency and process transparency. Our code is available at https://github.com/jayoungo/CloudMamba.
- 12 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 4704316
CloudMamba: An Uncertainty-Guided Dual-Scale Mamba Network for Cloud Detection in Remote Sensing Imagery
Cloud detection in remote sensing imagery is a fundamental, critical, and highly challenging problem. Existing deep learning-based cloud detection methods generally formulate it as a single-stage pixel-wise binary segmentation task with one forward pass. However, such single-stage approaches exhibit ambiguity and uncertainty in thin-cloud regions and struggle to accurately handle fragmented clouds and boundary details. In this paper, we propose a novel deep learning framework termed CloudMamba. To address the ambiguity in thin-cloud regions, we introduce an uncertainty-guided two-stage cloud detection strategy. An embedded uncertainty estimation module is proposed to automatically quantify the confidence of thin-cloud segmentation, and a second-stage refinement segmentation is introduced to improve the accuracy in low-confidence hard regions. To better handle fragmented clouds and fine-grained boundary details, we design a dual-scale Mamba network based on a CNN-Mamba hybrid architecture. Compared with Transformer-based models with quadratic computational complexity, the proposed method maintains linear computational complexity while effectively capturing both large-scale structural characteristics and small-scale boundary details of clouds, enabling accurate delineation of overall cloud morphology and precise boundary segmentation. Extensive experiments conducted on the GF1_WHU and Levir_CS public datasets demonstrate that the proposed method outperforms existing approaches across multiple segmentation accuracy metrics, while offering high efficiency and process transparency. Our code is available at https://github.com/jayoungo/CloudMamba.
- 12 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 22 (1545598X) , 1-5
Structural Representation-Guided GAN for Remote Sensing Image Cloud Removal
Optical remote sensing imagery is often compromised by cloud cover, making effective cloud removal techniques essential for enhancing the usability of such data. We designed a novel structural representation-guided GAN framework for cloud removal, in which structure and gradient branches are integrated into the network, helping the model focus on the structural representations of ground objects during image reconstruction. Different from previous methods that concentrate on recovering pixel information, we emphasize learning the structural information of remote sensing images. We then utilize error feedback to fuse features from the structural auxiliary branch, guiding the image reconstruction process. During the training phase, synthetic cloud images are used to supervise the optimization of the cloud removal network, while real cloud images are employed in an adversarial training manner for unsupervised learning to improve the generalization ability of the network. Additionally, multi-temporal revisit images from remote-sensing satellites are employed as auxiliary inputs, aiding the network to remove thick clouds reliably. We evaluated our framework on a dataset derived from SEN12MS-CR, and the proposed method outperformed classical cloud removal methods in both objective performance and subjective visual quality. Furthermore, compared to other methods, our approach achieved superior cloud removal results on real images.
- 12 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 22 (1545598X) , 1-5
Structural Representation-Guided GAN for Remote Sensing Image Cloud Removal
Optical remote sensing imagery is often compromised by cloud cover, making effective cloud removal techniques essential for enhancing the usability of such data. We designed a novel structural representation-guided GAN framework for cloud removal, in which structure and gradient branches are integrated into the network, helping the model focus on the structural representations of ground objects during image reconstruction. Different from previous methods that concentrate on recovering pixel information, we emphasize learning the structural information of remote sensing images. We then utilize error feedback to fuse features from the structural auxiliary branch, guiding the image reconstruction process. During the training phase, synthetic cloud images are used to supervise the optimization of the cloud removal network, while real cloud images are employed in an adversarial training manner for unsupervised learning to improve the generalization ability of the network. Additionally, multi-temporal revisit images from remote-sensing satellites are employed as auxiliary inputs, aiding the network to remove thick clouds reliably. We evaluated our framework on a dataset derived from SEN12MS-CR, and the proposed method outperformed classical cloud removal methods in both objective performance and subjective visual quality. Furthermore, compared to other methods, our approach achieved superior cloud removal results on real images.
- Open Access
- 23 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 17 (19391404) , 15206-15221
Weakly Supervised Adversarial Training for Remote Sensing Image Cloud and Snow Detection
Cloud and snow detection in remote sensing images has advanced significantly with the aid of deep learning methods. However, deep learning methods necessitate a large quantity of labeled data, which consumes a substantial amount of human and material resources. Numerous studies have focused on weakly supervised methods to reduce the workload of annotation, but the majority of these methods concentrate on cloud detection and involve snow detection only infrequently. In this paper, we propose a novel weakly supervised cloud and snow detection (WCSD) method. Under the guidance of the remote sensing imaging mechanism, we design generative adversarial networks (GAN) to generate cloud and snow images and pseudo labels for training detection networks. The proposed method can generate clouds of different states and reproduce snow's texture. For both the cloud GAN model and snow GAN model, with only imagelevel annotation training supervision, the models produce both pixel-level cloud/snow reflectance and cloud opacity to obtain the generated remote sensing images and corresponding pseudo labels. Compared to other weakly supervised methods, our method achieves superior cloud and snow detection performance.
- 23 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 17 (19391404) , 15206-15221
Open AccessWeakly Supervised Adversarial Training for Remote Sensing Image Cloud and Snow Detection
Cloud and snow detection in remote sensing images has advanced significantly with the aid of deep learning methods. However, deep learning methods necessitate a large quantity of labeled data, which consumes a substantial amount of human and material resources. Numerous studies have focused on weakly supervised methods to reduce the workload of annotation, but the majority of these methods concentrate on cloud detection and involve snow detection only infrequently. In this paper, we propose a novel weakly supervised cloud and snow detection (WCSD) method. Under the guidance of the remote sensing imaging mechanism, we design generative adversarial networks (GAN) to generate cloud and snow images and pseudo labels for training detection networks. The proposed method can generate clouds of different states and reproduce snow's texture. For both the cloud GAN model and snow GAN model, with only imagelevel annotation training supervision, the models produce both pixel-level cloud/snow reflectance and cloud opacity to obtain the generated remote sensing images and corresponding pseudo labels. Compared to other weakly supervised methods, our method achieves superior cloud and snow detection performance.